Skip to content

Summary of 14_LightGBM_SelectedFeatures

<< Go back

LightGBM

  • n_jobs: 6
  • objective: multiclass
  • num_leaves: 95
  • learning_rate: 0.2
  • feature_fraction: 0.5
  • bagging_fraction: 1.0
  • min_data_in_leaf: 10
  • metric: custom
  • custom_eval_metric_name: f1
  • num_class: 6
  • explain_level: 1

Validation

  • validation_type: kfold
  • k_folds: 5
  • shuffle: True
  • stratify: True
  • random_seed: 42

Optimized metric

f1

Training time

39.5 seconds

Metric details

0 1 2 3 4 5 accuracy macro avg weighted avg logloss
precision 0.970963 0.968872 0.966904 0.990241 0.995068 0.998444 0.981485 0.981749 0.981568 0.0617018
recall 0.99147 0.956466 0.968878 0.99177 0.992892 0.980642 0.981485 0.980353 0.981485 0.0617018
f1-score 0.981109 0.962629 0.96789 0.991005 0.993979 0.989463 0.981485 0.981013 0.981485 0.0617018
support 2462 1562 1960 1944 1829 1963 0.981485 11720 11720 0.0617018

Confusion matrix

Predicted as 0 Predicted as 1 Predicted as 2 Predicted as 3 Predicted as 4 Predicted as 5
Labeled as 0 2441 8 1 11 0 1
Labeled as 1 4 1494 61 0 3 0
Labeled as 2 15 40 1899 2 2 2
Labeled as 3 13 0 2 1928 1 0
Labeled as 4 9 0 1 3 1816 0
Labeled as 5 32 0 0 3 3 1925

Learning curves

Learning curves

Permutation-based Importance

Permutation-based Importance

Confusion Matrix

Confusion Matrix

Normalized Confusion Matrix

Normalized Confusion Matrix

ROC Curve

ROC Curve

Precision Recall Curve

Precision Recall Curve

<< Go back